USGS – Menlo Park's program for using high-rate, real-time GPS for earthquake rapid assessment in the San Francisco Bay area Abstract


  • The USGS office in Menlo Park has been working with partners to upgrade existing PBO's GPS stations and to install new GPS stations in the San Francisco Bay area with the primary focus on the Hayward fault. Our goal is to have on-hand high-rate GPS data (1 sample per second) with low latency obtained through a robust and reliable telemetry system. With reliable data following a large and damaging earthquake, we expect that the GPS measurements of position changes will augment seismic data to provide better estimate of the extent and the distribution of fault rupture. These products can be part of earthquake rapid assessment and provide information on the potential distribution of damage to infrastructure. It is also anticipated that a large earthquake on the Hayward fault will have substantial, postseismic slip for which the real-time, high rate GPS measurement can monitor the deformation. In constructing this network, we have worked with UNAVCO to harden telemetry links to PBO stations by converting from cell-phone modems, that depend upon commercial infrastructure, to radio links that take advantage of the existing USGS's Northern Californian Seismic Network's telemetry infrastructure. In addition, in the past year, UNAVCO installed eight new GPS stations for the USGS throughout the Bay area to complement the existing PBO and UCB stations. Our network uses line-of-sight, radio telemetry to various nodes which have direct Internet connection to USGS at Menlo Park using a set of micro-wave transponders. To monitor the operation of the telemetry of our GPS network, we have set up a website that provides quick and intuitive plots displaying the state of health of the telemetry and GPS receiver. A sister website provides plots of station displacements estimated using one of two computer programs that ingest streams of raw GPS data and produce estimates of position at each GPS site. One of the pieces of software is commercial while the other piece is open source. Currently we only are producing differential positions, but with the commercial software we do have an option of estimating positions using a precise, point positioning algorithm. We are currently testing both computer programs with the goal of producing reliable estimates of positions with minimal drift and outliers. The position estimates will be ingested into other programs to continuously evaluate the status of the deformation within our network; when deformation exceeds a threshold, another set of programs will quickly estimate the extent and distribution of slip should a large earthquake occur within the SF Bay area.


publication date

  • 2012

presented at event